Duplicating an app runtime feature for feature and bug for bug is difficult in a capital intensive sort of way. Even companies that can afford the expenditure, primarily Microsoft, decided to get out of that game because that level of expenditure comes with its own opportunity costs even at their level.
And there are alternatives: WebKit and Mozilla, one of those is the exclusive web renderer for iPhones and iPads which is not an insignificant amount of hardware in the world.
Google puts the work in. It’s really just that, and it’s their private prerogative to put that work in if they so choose, where others choose not to. The diversity of rendering engines isn’t great compared to say 15 years ago, but it’s not so bad that it’s a total monoculture in the way IE6 was. Actually thinking about it, more people are using more and different browser rendering engines than back then because Chromium-based browsers don’t have the total dominance that IE6 did, and there’s a lot more people on the web than there used to be. We had diversity, but we’re talking iCab, pre-WebKit OmniWeb, Opera on Presto, and yeah, Trident’s not really around anymore, but it was there, and so was Gecko and WebKit.
At least you don’t need a browser that supports ActiveX controls for regular websites anymore.
My problem with Chrome dominance is that it's both a dominance of a single company (since Chrome is by far the dominant "fork" of Chromium, and owned by an advertising corporation), and a dominance of a single browser engine. The latter is IMO a really big threat, as it leads to browser engine monoculture. Monoculture leads to a single point of failure, and in this case, a loss of knowledge about how to build a browser engine that's "the websites that are actually out there"-compliant. Much more relies on the web now, so it's paramount that we have a useful alternative browser engine implementation.